Softechinfra
Business

UK Closes AI Chatbot Loophole in Online Safety Act

PM Starmer extends the Online Safety Act to ChatGPT, Gemini, and Copilot — banning infinite scroll and AI chatbot access for children across platforms.

Softechinfra EditorialSoftechinfra Editorial
March 21, 20268 min read
UK Closes AI Chatbot Loophole in Online Safety Act

The United Kingdom has moved decisively to close a gap that regulators had flagged since the Online Safety Act first passed: general-purpose AI chatbots were not covered. PM Keir Starmer's government announced on February 16 that ChatGPT, Google Gemini, and Microsoft Copilot would now fall within the Act's scope.

What Changed

1
Scope Extension
ChatGPT, Gemini, and Copilot classified as regulated services under the Online Safety Act, triggering Ofcom oversight.
2
Minimum Age Enforcement
Platforms must implement verifiable age checks — not mere self-declaration.
3
Harmful Feature Restrictions
Infinite scrolling, autoplay, and algorithmic recommendation must be restricted for child accounts.
4
VPN Circumvention Controls
Platforms must take reasonable measures to prevent VPN-based age restriction bypass.

Implications for Product Teams

Any product that incorporates a third-party AI chatbot now inherits potential compliance obligations under the Online Safety Act if it has UK users. Product teams building for under-18 audiences can look at how we approached this with a client platform, which enforces age gates, filtered prompts, and learning-focused outputs by default.

  • Audit every AI-powered conversational feature for potential child-facing exposure
  • Review your age verification flow — self-reported date of birth is no longer sufficient
  • Disable engagement-maximising features for users under 18
  • Ensure vendor contracts with AI API providers include compliance obligations
  • Monitor Ofcom guidance — implementation codes are still being published
Context: The Online Safety Act imposes fines of up to 10% of global annual turnover for non-compliance.

Our web development and custom software teams are already advising clients building UK-facing products. Talk to us before your next release.

For a wider founder take on AI regulation — and why the UK approach is likely to become a template for other jurisdictions — see viveksinra.com.

Tags:
RegulationAI SafetyOnline Safety ActUK PolicyChildren Online
Share this post:
Softechinfra Editorial

Softechinfra Editorial

Articles published by Softechinfra without a personal byline.

How we pick topics, use AI, check claims and correct mistakes: our editorial policy.